Re: [HOT] Most "active" HOT projects since March 2014
> On 23 Mar 2015, at 20:35, Pierre Béland wrote: > > The number of changesets is also a gross estimate of the contributor efforts. > How to compare a contributor that adds 100 buildings, one changeset for each > (we see this often with new contributors) with a contributor that has a few > thousands objects in one edit. To address this I just added one additional tab to the spreadsheet — HOT projects ranked by the number of changes made to the map. (This corresponds to the “num_changes” measure in OSM changeset stats.) https://docs.google.com/spreadsheets/d/12TjMDsgPoEQadiUWbpKjotTMYKPo5NY4W7EDipHnE7Y/edit?usp=sharing #MapLesotho (project 599) at the top, with a wide margin... Thanks again for the feedback! m. ___ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot
Re: [HOT] Most "active" HOT projects since March 2014
I just updated the spreadsheet — the user numbers were inflated thanks to a subtle bug in my code with big consequences… they now more closely track total participant counts in the HOT tasking manager. (Many thanks to DaCor and Ciaran Staunton who helped me track this down.) m. > On 23 Mar 2015, at 18:59, Martin Dittus wrote: > > Hallo all, > > Here are two basic rankings of HOT project activity in the last year, based > on map contributions in the past 12 months. Discussion on IRC suggested that > this data might be useful to others, so I’m sharing it with the list. > > HOT projects ranked by edit activity from Mar 2014 - Feb 2015 (inclusive): > https://docs.google.com/spreadsheets/d/12TjMDsgPoEQadiUWbpKjotTMYKPo5NY4W7EDipHnE7Y/edit?usp=sharing > > The first tab shows a ranking of all tasking manager projects by number of > changesets, the second by number of contributors. > > This data is not yet captured by the tasking manager: I’m only looking at > participation that actually resulted in changes on the map. To this purpose I > identify changesets in the OSM edit history that were tagged with a HOT > project id. > > I’m curious what other people can read from this data, e.g. whether it > matches your intuitions of popular HOT activities. > > Of course these kinds of rankings are of dubious utility — projects are > rarely directly comparable in their scope, and larger activations may be > structured in all kinds of ways. Same goes for changesets as a metric. > > Does anyone have good suggestions for how one could meaningfully group the > hundreds of HOT projects? E.g. I could aggregate stats for all > #MissingMaps/#MapLesotho/Ebola Outbreak projects, are there similar tags I > should be looking for? Are these actually meaningful distinctions when > looking at contribution outcomes? > > Any requests for other stats? I’d be happy to produce more, provided it’s > feasible. No promises :) > > > Greetings from London, > Martin Dittus > > --- > > If you’re interested in the details — there are a number of challenges in > producing such data, and some caveats. Mostly it has to do with the > unstructured nature of changeset comments. > > I’m likely undercounting: automated iD changeset comments were only > introduced in mid-August 2014, see > http://www.openstreetmap.org/user/mcld/diary/24123 — I’m not identifying > earlier contributions that have not been tagged manually. > > I’m likely undercounting: HOT projects use multiple tagging conventions for > changeset comments, most prominently #hotosm-project-938 and #hotosm-task-907 > but also #hotosm-Ebola-892 and similar. I'm being quite lenient in what I > expect but may not catch all projects. > > I also noticed the tag #hotosm-cap103 which is not actually referring to a > project ID, that's an activation ("Projet Cap103”). > > And — I don’t have project titles for projects that aren’t public any longer. > ___ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot
Re: [HOT] Most "active" HOT projects since March 2014
These heatmaps are interesting. Neis changeset online map shows intensity of mapping only with the outset of the bbox. Such heatmap could show the intensity of contribution in various areas based on the number of objects. To use this functionality for an online map with Openlayers or similar javascript tool, we would surely need to aggregate the data. Otherwise, I dont see with a project like the Ebola outbreak, how we could manage nearly 16 millions objects to represent on the Heatmap. Pierre De : Martin Dittus À : Pierre Béland Cc : hot Envoyé le : Lundi 23 mars 2015 18h32 Objet : Re: [HOT] Most "active" HOT projects since March 2014 Thanks for the detailed note Pierre! We seem to be on the same page — such stats are wonderfully misleading. I’ve tried grouping HOT projects in the past based on the available metadata and have indeed encountered all the obstacles you mention. Part of my motivation to share this was also to gently nudge project creators to tag more consistently :) Such metadata can be very useful for large-scale evaluations and visualisations. Pierre Béland wrote: > > The best way to reflect the contributor activity is to go through the history > and extract changesets corresponding to a bbox closed to the each Task > Manager extent and corresponding to the period of the project. Yeah I’ve done that in the past on an older OSM history dump, that’s how this world map of HOT contributions was created: https://twitter.com/dekstop/status/565211831720235009 Although that image actually visualises individual edits, not just changesets. m. ___ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot
Re: [HOT] Most "active" HOT projects since March 2014
Thanks for the detailed note Pierre! We seem to be on the same page — such stats are wonderfully misleading. I’ve tried grouping HOT projects in the past based on the available metadata and have indeed encountered all the obstacles you mention. Part of my motivation to share this was also to gently nudge project creators to tag more consistently :) Such metadata can be very useful for large-scale evaluations and visualisations. Pierre Béland wrote: > > The best way to reflect the contributor activity is to go through the history > and extract changesets corresponding to a bbox closed to the each Task > Manager extent and corresponding to the period of the project. Yeah I’ve done that in the past on an older OSM history dump, that’s how this world map of HOT contributions was created: https://twitter.com/dekstop/status/565211831720235009 Although that image actually visualises individual edits, not just changesets. m. ___ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot
Re: [HOT] Most "active" HOT projects since March 2014
Hi Martin, It is uneasy to build a statistic that would reflect the various activities. Coordination through the Task manager is only one part of the mapping done for projects. It might be ok a mapathon where you worked only coordinating through the Task Manager. Aslo, Hashtags are not always added to the changeset comment. The best way to reflect the contributor activity is to go through the history and extract changesets corresponding to a bbox closed to the each Task Manager extent and corresponding to the period of the project. The number of changesets is also a gross estimate of the contributor efforts. How to compare a contributor that adds 100 buildings, one changeset for each (we see this often with new contributors) with a contributor that has a few thousands objects in one edit. About the grouping, you have main activities like the Ebola outbreak. At the same time you can have various mapping projects that contribute to this action. For example, a missing map party on Ebola, where should we classify it. Also, the Task Manager represents only one part of contributions. As an example, for the Ebola outbreak the experienced mappers move around and correct / enhance various elements of the map without coordinating through the Tasking manager. Pierre De : Martin Dittus À : hot Envoyé le : Lundi 23 mars 2015 14h59 Objet : [HOT] Most "active" HOT projects since March 2014 Hallo all, Here are two basic rankings of HOT project activity in the last year, based on map contributions in the past 12 months. Discussion on IRC suggested that this data might be useful to others, so I’m sharing it with the list. HOT projects ranked by edit activity from Mar 2014 - Feb 2015 (inclusive): https://docs.google.com/spreadsheets/d/12TjMDsgPoEQadiUWbpKjotTMYKPo5NY4W7EDipHnE7Y/edit?usp=sharing The first tab shows a ranking of all tasking manager projects by number of changesets, the second by number of contributors. This data is not yet captured by the tasking manager: I’m only looking at participation that actually resulted in changes on the map. To this purpose I identify changesets in the OSM edit history that were tagged with a HOT project id. I’m curious what other people can read from this data, e.g. whether it matches your intuitions of popular HOT activities. Of course these kinds of rankings are of dubious utility — projects are rarely directly comparable in their scope, and larger activations may be structured in all kinds of ways. Same goes for changesets as a metric. Does anyone have good suggestions for how one could meaningfully group the hundreds of HOT projects? E.g. I could aggregate stats for all #MissingMaps/#MapLesotho/Ebola Outbreak projects, are there similar tags I should be looking for? Are these actually meaningful distinctions when looking at contribution outcomes? Any requests for other stats? I’d be happy to produce more, provided it’s feasible. No promises :) Greetings from London, Martin Dittus --- If you’re interested in the details — there are a number of challenges in producing such data, and some caveats. Mostly it has to do with the unstructured nature of changeset comments. I’m likely undercounting: automated iD changeset comments were only introduced in mid-August 2014, see http://www.openstreetmap.org/user/mcld/diary/24123 — I’m not identifying earlier contributions that have not been tagged manually. I’m likely undercounting: HOT projects use multiple tagging conventions for changeset comments, most prominently #hotosm-project-938 and #hotosm-task-907 but also #hotosm-Ebola-892 and similar. I'm being quite lenient in what I expect but may not catch all projects. I also noticed the tag #hotosm-cap103 which is not actually referring to a project ID, that's an activation ("Projet Cap103”). And — I don’t have project titles for projects that aren’t public any longer. ___ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot ___ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot
[HOT] Most "active" HOT projects since March 2014
Hallo all, Here are two basic rankings of HOT project activity in the last year, based on map contributions in the past 12 months. Discussion on IRC suggested that this data might be useful to others, so I’m sharing it with the list. HOT projects ranked by edit activity from Mar 2014 - Feb 2015 (inclusive): https://docs.google.com/spreadsheets/d/12TjMDsgPoEQadiUWbpKjotTMYKPo5NY4W7EDipHnE7Y/edit?usp=sharing The first tab shows a ranking of all tasking manager projects by number of changesets, the second by number of contributors. This data is not yet captured by the tasking manager: I’m only looking at participation that actually resulted in changes on the map. To this purpose I identify changesets in the OSM edit history that were tagged with a HOT project id. I’m curious what other people can read from this data, e.g. whether it matches your intuitions of popular HOT activities. Of course these kinds of rankings are of dubious utility — projects are rarely directly comparable in their scope, and larger activations may be structured in all kinds of ways. Same goes for changesets as a metric. Does anyone have good suggestions for how one could meaningfully group the hundreds of HOT projects? E.g. I could aggregate stats for all #MissingMaps/#MapLesotho/Ebola Outbreak projects, are there similar tags I should be looking for? Are these actually meaningful distinctions when looking at contribution outcomes? Any requests for other stats? I’d be happy to produce more, provided it’s feasible. No promises :) Greetings from London, Martin Dittus --- If you’re interested in the details — there are a number of challenges in producing such data, and some caveats. Mostly it has to do with the unstructured nature of changeset comments. I’m likely undercounting: automated iD changeset comments were only introduced in mid-August 2014, see http://www.openstreetmap.org/user/mcld/diary/24123 — I’m not identifying earlier contributions that have not been tagged manually. I’m likely undercounting: HOT projects use multiple tagging conventions for changeset comments, most prominently #hotosm-project-938 and #hotosm-task-907 but also #hotosm-Ebola-892 and similar. I'm being quite lenient in what I expect but may not catch all projects. I also noticed the tag #hotosm-cap103 which is not actually referring to a project ID, that's an activation ("Projet Cap103”). And — I don’t have project titles for projects that aren’t public any longer. ___ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot